AI nervous system that senses, decides, and acts
in sync with every organization.
It reads the signals every organization already produces: Slack, calendars, PTO, 1:1s, feedback, recognition, pulse. It decides what matters and acts on it, before anyone has to ask.
Founder-led onboarding · Live in under a day
May 2026
launched, live in production
17,000
LinkedIn followers in two weeks, zero paid spend
20+
enterprise calls booked, inbound
It senses. It decides. It acts.
Headcount reads PTO, pulse, feedback, 1:1s, recognition, goals, and meeting transcripts continuously, decides what actually matters, and acts on it before you have to ask. Ask a question and get the full picture, or let it reach out first.
Ask anything about your team
Team health, burnout risk, recognition gaps, who's out this week. Ask in plain English and the agent synthesizes nine signal streams into a grounded answer.
- Per-person burnout breakdown by dimension, with the signals that drove the score
- Coverage risk for the week ahead: who's out, who's covering, where workload concentrates
- Manager Effectiveness percentile vs your peers across cadence, recognition, follow-through, and sentiment
- Recognition gaps by IC, with what specifically to recognize from the last two weeks
Works where you already work
Slack, Teams, Google Chat, wherever your team already messages. The agent lives there, not in a new app you have to install.
- Slack today. @mention in channels or DM the agent directly.
- Microsoft Teams and Google Chat: coming next.
- Web: full chat panel built into the Headcount dashboard.
- Employees respond to pulse check-ins and give feedback without leaving chat.
It finds the patterns you'd never connect by hand
The agent computes burnout, recognition gaps, and pulse trends continuously in the background. When you ask, the synthesis is ready. You skip the homework.
- Continuous burnout score per person, not a survey-once-a-quarter snapshot
- Pulse free-text synthesis: spots when three people are flagging the same thing in different words
- Recognition gap detection with the specific work to call out, not just names on a list
- Privacy-respecting signals only, with a k-anonymity floor on every aggregate. No message reading, no surveillance.
It doesn't just tell you. It does it.
'Help me close my recognition gap this week' and the agent finds the right people, surfaces what specifically to call out, drafts the message in your voice, and sends on confirmation. Recall.ai joins your 1:1s and standups and turns the transcript into draft pulse responses for every participant.
- Recognition with evidence: the agent supplies the 'what to call out', not just the name
- 1:1 scheduling with topics, action items, and notes in one conversation
- Meeting-linked Pulse: bot joins your meeting and drafts pulse responses from the transcript
- Voice-adaptive drafts. Every action confirmed before it runs. Every outcome captured so the next decision is sharper.
Goals that draft themselves from how the team is actually doing
Ask the agent to set a goal for someone and it grounds the objective and key results in real signals: workload, recent feedback, 1:1 cadence, recognition. You edit, confirm, and it links to the team goal it ladders up to.
- AI-drafted objectives with 2 to 5 weighted key results, grounded in each person's real context, not a blank template
- Status derives itself from pace and staleness: a goal that quietly stops moving flags itself at-risk, with the reason
- One-level alignment: an individual goal links to the team goal it supports. No cascading OKR tree to maintain
- Every AI suggestion is a draft you confirm. Nothing is written to a goal without you
Six compounding moats.
Every deployment gets harder to leave, on purpose. Here's what compounds underneath it.
Behavioral graph compounds
Every deployment builds an organization-specific graph that gets more valuable every week. Switching tools means rebuilding years of organizational intelligence.
Every interaction trains the system
Accept, edit, and ignore continuously refine what gets recommended to each manager and each organization.
Organization-first architecture
Models relationships, reporting lines, and team dynamics, not isolated conversations.
Embedded into weekly rituals
Lives inside 1:1s, weekly briefs, recognition, and goal planning. Removing it means changing how managers already work.
The dataset compounds passively
Slack, meetings, and calendars continuously update the graph. No one has to remember to ask.
Grounded, defensible scoring
Burnout is scored against a clinical measure, per person, over time. Not a generic opinion.
From zero to a live nervous system in one afternoon.
No migration. No training. No six week rollout. Connect your tools and the loop starts running.
Connect your tools
Add Headcount to Slack or Teams. Invite your team. We request only the permissions we need.
The graph starts building
Headcount fuses time off, 1:1 history, feedback, pulse responses, recognition, and meeting transcripts into a living, per-organization model that sharpens every week.
Talk to your agent
Ask questions, get alerts, let it handle things. Everything happens in Slack or Teams. No new app. No dashboard. Just talk.
Cheaper than the hire. Everywhere the hire can't be.
One chief of staff is about $8,000 a month. Headcount covers your whole team for $19 a seat.
A human chief of staff
one person. one perspective.
one salary, one set of hands
Still hiring for this?
Headcount
your AI chief of staff. every person, every week.
$15/seat/month billed annually. $19/seat when billed monthly.
We'll set you up in under a week
Enterprise
For larger organizations
Customtailored to your needs
Volume discounts available for 200+ seats
Everything you're wondering.
Workday manages records. Rippling manages admin. Headcount manages the company.
Ask about your team. Let the agent handle the rest.
Founder-led onboarding · Live in under a day · 7-day pilot activation